Corpus linguistic tools are used by an increasingly diverse range of academics with no previous e... more Corpus linguistic tools are used by an increasingly diverse range of academics with no previous expertise in linguistics to support the analysis of their language data. We are particularly interested in the ways in which corpus linguistic tools can aid anthropologists. Our corpus-assisted discourse analytic investigation into religious tolerance therefore uses language-based anthropological data. This data emanates from an ethnographic survey (carried out in southwest Nigeria in 2012-13 and involving more than 2800 participants) gathered as part of the larger research project: ‘Knowing each other: everyday religious encounters, social identities and tolerance in southwest Nigeria’ (henceforth ‘KEO’). This project examines the coexistence of Islam, Christianity and traditional practice in the Yoruba-speaking parts of Nigeria. It focusses on how people encounter other religions and how their identity is shaped by this encounter. It also posits that exploring the beliefs and attitudes of participants about these encounters with the religious ‘others’ can contribute to achieving these aims. Our corpus differs considerably from what usually gets referred to as ‘corpora’ in corpus linguistics for it consists of the answers to all 60 open-ended questions in the above-mentioned survey. These questions asked respondents either to discuss their own or family members’ experiences of inter-religious encounter, or their views on hypothetical scenarios relating to inter-religious encounter. In the part of the questionnaire focusing on their own experiences, respondents were asked to explain (if applicable) why they or family members had changed their religion, any attempts others had made to convert them, and the reasons they felt people in general converted. They were also asked for their views on inter-religious marriage and their experiences and views on religious differences between children and parents. Another part of the survey asked them about their experiences of participating in practices associated with other religions or Yoruba traditions, including hypothetical scenarios such as whether they would visit an alfa (Islamic cleric) or pastor if they had a problem, whether they would allow family members of a different religion to attend a family celebration, and about how they would accommodate family members and friends of different religions at social events. Finally, they were asked what they liked or respected about Islam, Christianity and Yoruba customs, how they would advise religious leaders to behave towards one another, their experiences of religious conflict and how they would suggest such conflict could be prevented. Our corpus is therefore thematic, for it captures discourse about interreligious encounters in Yorubaland in South West Nigeria and we use it to ‘investigate cultural attitudes expressed through language’ (Hunston 2002: 13-14). Insofar as the reasons for the survey were clear and the data collection was rigorous, we argue that the KEO corpus is homogeneous. The overall corpus is ‘bilingual’ in that some participants answered in English and others in Yoruba. However, we use the English component of the KEO corpus, i.e. all the answers in English (approximately 300,000 words). The KEO corpus evidently falls under the category of ‘ad hoc’ specialised corpora (specialised corpora are particularly relevant for a range of social scientists for whom a corpus of general English is not relevant) and is relatively small owing to the large corpora other corpus-driven research, for example, utilises. We distinguish answers to the survey questions provided by Muslim and Christian participants (i.e. two subcorpora) as a means to examine the discursive choices they make when discussing each other. First, we compare the ICE Nigeria corpus (insofar as the English used by our participants is South West Nigerian English) with our KEO English corpus (e.g. comparison of ‘key lemma’ lists to explore the ‘aboutness’ of our corpus and to select a group of words for further study). Second, we compare our two subcorpora (all answers by Muslim and Christian participants). We use a range of corpus outputs for each subcorpus (e.g. word frequencies) to give us an initial insight into the difference and/or lack of difference between the two religious groups under scrutiny, and comment on whether these are meaningful. Third, we delve deeper into the language used by Christians and Muslims to discuss the religious ‘other’ with a view to gain further insight into what Muslims and Christians’ perception of themselves and each other. For example, patterns associated with the words ‘Islam’ and ‘Muslim’ on the one hand and ‘Christianity’ and ‘Christian’ on the other hand are examined (e.g. using concordance and collocation lists to examine the contexts of these specific words in the two sucbcorpora). To conclude, our corpus is clearly atypical for it captures data which does not fall neatly under what is usually understood to be a ‘corpus’ by linguists using a corpus-based paradigm. Our methodological approach therefore raises a range of timely questions and issues for social scientists wishing to use corpus tools in their research. We will therefore also ask what kinds of new lines of enquiry, if any, corpus-assisted discourse analytic methodology can suggest for anthropologists.
References Hunston, S. 2002. Corpora in applied linguistics. Cambridge: Cambridge University Press.
Our paper will explore the ways in which religious identities are negotiated in a setting charact... more Our paper will explore the ways in which religious identities are negotiated in a setting characterised by religious diversity and proximity: Yorubaland in South West Nigeria. We will explore how interreligious relationships are discursively constructed in extensive survey data (2,819 respondents in total) collected as part of an anthropological project focussing on the coexistence of Islam, Christianity and traditional practice in Yoruba-speaking parts of southwest Nigeria: ‘Knowing each other: Everyday religious encounters, social identities and tolerance in southwest Nigeria’. Corpus tools and techniques will be used to examine the 1,535 questionnaires filled in English, particularly answers to open-ended questions (our corpus). The premise is that by exploring discursive choices made by Christian and Muslim respondents in this corpus, we can gain insights into Yoruba Muslims and Christians’ perception of themselves and each other and their experiences of inter-religious encounters. Owing to the focus of the paper, two sub-corpora of the above-mentioned corpus were compiled: one with all answers by respondents of Muslim faith and another with all answers by respondents of Christian faith. We will use four-grams for each of these corpora to show how corpus-assisted investigations into phraseology have helped us gain insights into the data which traditional anthropological methods alone would not have allowed. Our findings will concern, for example, the specific boundaries our Christian and Muslim respondents draw around their religious behaviour and their shared understanding of religion.
The three books reviewed in this article (Osborn, Our New Husbands Are Here; Jean-Baptiste, Conju... more The three books reviewed in this article (Osborn, Our New Husbands Are Here; Jean-Baptiste, Conjugal Rights; Burrill, States of Marriage) make an important contribution to Africa's social and political history. Defying explorations of marriage as a product of the political economy and avoiding a consideration of marriage simply as an example of social or legal transformation, Emily Lynn Osborn, Rachel Jean-Baptiste and Emily Burrill focus on marriage, sexual relations and the household as important and contested categories of historical production. Their emphasis on the implication of marriage in wider social relationships resonates with historical-anthropological reconstructions of precolonial and colonial forms of gendered agency. It also reveals marriage as a vantage point for research that transcends conceptual boundaries between the private and the political, and by implication also boundaries between different fields of historical inquiry. The authors' shared emphasis on marriage as a 'messy engagement', embedded in family relationships, social obligations and political interests, makes for fascinating reading.
Uploads
Our corpus differs considerably from what usually gets referred to as ‘corpora’ in corpus linguistics for it consists of the answers to all 60 open-ended questions in the above-mentioned survey. These questions asked respondents either to discuss their own or family members’ experiences of inter-religious encounter, or their views on hypothetical scenarios relating to inter-religious encounter. In the part of the questionnaire focusing on their own experiences, respondents were asked to explain (if applicable) why they or family members had changed their religion, any attempts others had made to convert them, and the reasons they felt people in general converted. They were also asked for their views on inter-religious marriage and their experiences and views on religious differences between children and parents. Another part of the survey asked them about their experiences of participating in practices associated with other religions or Yoruba traditions, including hypothetical scenarios such as whether they would visit an alfa (Islamic cleric) or pastor if they had a problem, whether they would allow family members of a different religion to attend a family celebration, and about how they would accommodate family members and friends of different religions at social events. Finally, they were asked what they liked or respected about Islam, Christianity and Yoruba customs, how they would advise religious leaders to behave towards one another, their experiences of religious conflict and how they would suggest such conflict could be prevented. Our corpus is therefore thematic, for it captures discourse about interreligious encounters in Yorubaland in South West Nigeria and we use it to ‘investigate cultural attitudes expressed through language’ (Hunston 2002: 13-14). Insofar as the reasons for the survey were clear and the data collection was rigorous, we argue that the KEO corpus is homogeneous. The overall corpus is ‘bilingual’ in that some participants answered in English and others in Yoruba. However, we use the English component of the KEO corpus, i.e. all the answers in English (approximately 300,000 words). The KEO corpus evidently falls under the category of ‘ad hoc’ specialised corpora (specialised corpora are particularly relevant for a range of social scientists for whom a corpus of general English is not relevant) and is relatively small owing to the large corpora other corpus-driven research, for example, utilises.
We distinguish answers to the survey questions provided by Muslim and Christian participants (i.e. two subcorpora) as a means to examine the discursive choices they make when discussing each other. First, we compare the ICE Nigeria corpus (insofar as the English used by our participants is South West Nigerian English) with our KEO English corpus (e.g. comparison of ‘key lemma’ lists to explore the ‘aboutness’ of our corpus and to select a group of words for further study). Second, we compare our two subcorpora (all answers by Muslim and Christian participants). We use a range of corpus outputs for each subcorpus (e.g. word frequencies) to give us an initial insight into the difference and/or lack of difference between the two religious groups under scrutiny, and comment on whether these are meaningful. Third, we delve deeper into the language used by Christians and Muslims to discuss the religious ‘other’ with a view to gain further insight into what Muslims and Christians’ perception of themselves and each other. For example, patterns associated with the words ‘Islam’ and ‘Muslim’ on the one hand and ‘Christianity’ and ‘Christian’ on the other hand are examined (e.g. using concordance and collocation lists to examine the contexts of these specific words in the two sucbcorpora).
To conclude, our corpus is clearly atypical for it captures data which does not fall neatly under what is usually understood to be a ‘corpus’ by linguists using a corpus-based paradigm. Our methodological approach therefore raises a range of timely questions and issues for social scientists wishing to use corpus tools in their research. We will therefore also ask what kinds of new lines of enquiry, if any, corpus-assisted discourse analytic methodology can suggest for anthropologists.
References
Hunston, S. 2002. Corpora in applied linguistics. Cambridge: Cambridge University Press.
Owing to the focus of the paper, two sub-corpora of the above-mentioned corpus were compiled: one with all answers by respondents of Muslim faith and another with all answers by respondents of Christian faith. We will use four-grams for each of these corpora to show how corpus-assisted investigations into phraseology have helped us gain insights into the data which traditional anthropological methods alone would not have allowed. Our findings will concern, for example, the specific boundaries our Christian and Muslim respondents draw around their religious behaviour and their shared understanding of religion.